LAUSR

INTRODUCTION Depression in people living with HIV/AIDS is associated with poor health outcomes. Despite this, assessment of depressive symptoms is not a routine clinical practice in the care of people with HIV in Colombia. One reason could be the lack of validated depression screening scales for this population. OBJECTIVE To test the reliability and construct validity of the 20- and 10-item-Center for Epidemiological Studies Depression Scale in patients attending an HIV clinic in Cali, Colombia. MATERIALS AND METHODS A non-random sample of 105 adults was enrolled. The 20 item-CES-D (CES-D-20) scale was administered twice: At baseline and 2-4 weeks later. We calculated the Cronbach's alpha coefficient and the intraclass correlation coefficient. In addition, we used an exploratory and confirmatory factorial analysis, as well as the item response theory to assess the validity of the scale. RESULTS Most participants were men (73%), with a mean age of 40 years, 53% of whom had not completed high school. Cronbach's coefficients were 0.92 and 0.94 at baseline and at the second interview, respectively. The intraclass correlation was 0.81 (95% CI: 0.72-0.88). Although all 20 items loaded distinctly in 4 factors, 5 items did not load as expected. The structure factor of the CES-D-20 was not confirmed, as 4 items had poor goodness of fit. The CES-D-10 appeared to perform better in this population. CONCLUSIONS These results support the reliability and validity of the CES-D-10 instrument to screen for depressive symptoms in people living with HIV in Colombia.